Discovery of disease specific biomarkers in human body fluids has become an important challenge in clinicalproteomics. Facing the increasing threat of degenerative and disabling diseases like cancer, cardiovascular, neurologicaland inflammatory diseases in large parts of the world population, there is an urgent need to improve early diagnostics. Inthis review we discuss possibilities and limitations connected to using mass spectrometry based proteomics in the searchfor novel biomarkers, with focus on multiple sclerosis as a typical representative for the large group of non-curable de-generative and disabling disease with the lack of specific tests for early diagnosis. Careful control of the pre-analyticalphase including sampling, storage and fractionation of samples, in addition to a thoroughly considered patient selection, isimportant in order to avoid false biomarkers to appear in the resulting mass spectra. Furthermore, advanced computationaltools are needed in order to discover potential biomar kers from the enormous data amounts generated by the mass spec-trometers. The development of such computer tools is a research field currently in the start phase and could prove to be abottle neck in the biomarker discovery the next years. Therefore, a rather detailed review of the most used computationaland pre-analytical methods is given in this review. Mass spectrometry based biomarker discovery is undoubtedly still inits early infancy. However, in light of the potential of this technology to provide deep coverage of the body fluid proteo-mes, it will certainly consolidate its role in developing molecular medicine into clinical practice.